A prospective, placebo-controlled, triple blind clinical trial was carried out in Pakistan to determine the effect of Lactobacillus GG on the course of acute diarrhea in hospitalized children. Forty children (mean age, 13 months) were enrolled and after rehydration received either oral Lactobacillus GG (n = 21) or placebo (n = 19) twice daily for 2 days, in addition to the usual diet. The clinical course of diarrhea was followed during the treatment period. Features on admission into the study groups were similar and were characterized by severe diarrhea, malnutrition and inappropriate management before presentation. Response was evident on Day 2 when the frequency of both vomiting and diarrhea was less in the Lactobacillus group. In those who had presented with acute nonbloody diarrhea (n = 32), the percentage of children with persistent watery diarrhea at 48 hours was significantly less in the Lactobacillus group: 31% vs. 75% (P < 0.01). No significant difference was observed by 48 hours in those presenting with bloody diarrhea. The relevance of this finding to the management of diarrhea in the tropics is discussed.